AI-based monitoring and optimization based on sensor data from injection molds (COMET K2)

The goal of this project is AI-based monitoring and optimization of algorithms based on sensor data from injection molds. Using a battery-powered injection mold sensor platform, an indoor localization system, and on-board artificial intelligence (AI) algorithms, the states of the injection mold (e.g., acceleration, distance, and dwell time) are provided as a real-time service to a cloud environment, independent of local conditions. In addition, more complex algorithms are developed in the cloud application for more detailed analysis of the recorded data and metadata. The number of processing cycles, the used sub-steps, processed materials and the used driving parameters (processing speed, pressure, …) strongly influence the mechanical stress of a tool. For innovative business models (e.g., remote monitoring as a service, pay-per-use of a component of the plant…), the reliable assessment of tool usage and stress including the identification of processing cycles with limited access to the plant is of high importance. Use of machine learning on edge systems with low computing power and/or with a small amount of training data plays a crucial role in the project. In several applications, data-driven methods with high computational complexity cannot be used or must be complemented by an edge system with low computational powe